private async void CurrentVideoFrame(ThreadPoolTimer timer)
{
//複数スレッドでの同時実行を抑制
if (!semaphore.Wait(0))
{
return;
}
try
{
//AIモデルのインプットデータは解像度224x224,BGRA8にする必要がある。
using (VideoFrame previewFrame = new VideoFrame(BitmapPixelFormat.Bgra8, 224, 224))
{
await this.mediaCapture.GetPreviewFrameAsync(previewFrame);
if (ModelGen != null && previewFrame != null)
{
ModelInput.image = Windows.AI.MachineLearning.ImageFeatureValue.CreateFromVideoFrame(previewFrame);
// Evaluate the model
ModelOutput = await ModelGen.EvaluateAsync(ModelInput);
// Convert output to datatype
var batchIdx = 0;
IDictionary <string, float> vectorImage = ModelOutput.classLabelProbs[batchIdx];
var scoreList = vectorImage.Values.ToList();
var labelList = vectorImage.Keys.ToList();
//IReadOnlyList<float> vectorImage = ModelOutput.Plus214_Output_0.GetAsVectorView();
//IList<float> imageList = vectorImage.ToList();
// Query to check for highest probability digit
var maxValue = scoreList.Max();
var maxIndex = scoreList.IndexOf(maxValue);
var label = labelList[maxIndex];
// Display the results on UI Thread.
var ignored = this.Dispatcher.RunAsync(Windows.UI.Core.CoreDispatcherPriority.Normal, () =>
{
string result = "";
//予測結果を表示
result = result + "Class: " + label + ", Prob: " + maxValue;
this.msgTbk.Text = result;
});
}
}
}
catch (Exception ex)
{
Debug.WriteLine(ex.Message);
}
finally
{
semaphore.Release();
}
}
public async Task <mobilenetv2Output> EvaluateAsync(mobilenetv2Input input)
{
binding.Bind("image", input.image);
var result = await session.EvaluateAsync(binding, "0");
var output = new mobilenetv2Output();
output.classLabel = result.Outputs["classLabel"] as TensorString;
output.classLabelProbs = result.Outputs["classLabelProbs"] as IList <IDictionary <string, float> >;
return(output);
}
